Hearing aids compensate for a patient's loss of hearing function by enhancing ambient acoustic sounds. This is done via detecting ambient acoustic signals, processing the signals according to a patient specific prescription, and delivering the processed signals to the patient in a manner that the patient perceives as sound. Hearing aids are often categorized into one of two types, namely conventional and implantable hearing aids. Implantable hearing aids may be further categorized into fully implantable devices and semi-implantable devices.
Conventional hearing aids typically include a microphone, amplifier, signal processor, and speaker and are worn behind the ear and/or in the ear canal of the patient. Semi-implantable hearing aids typically include, a microphone, amplifier, signal processor, and transmitter that are externally located and inductively transmit a processed audio signal to an implanted receiver and transducer. Fully implantable-hearing aids, on the other hand, locate the microphone, amplifier, signal processor, and transducer subcutaneously below a patients skin, e.g. typically in the mastoid process and/or middle ear cavity.
Unfortunately, hearing aid devices, such as those described above, are often subject to feedback oscillation, e.g. resonant phenomenon due to re-amplification of feedback signals having a net phase of zero degrees. In conventional devices, the feedback is most often provided over a feedback path leading through the air to the microphone where it is re-amplified by the amplifier located downstream from the microphone. In implanted devices, the feedback may be provided over different propagation paths to the microphone and amplifier, such as via the eardrum and middle ear canal or the bones and/or other parts of the skull. In this regard, feedback signals are reintroduced to the microphone of the hearing aid where they may be re-amplified again by the amplifier to create an oscillation. When feedback signals, audible or not, oscillate through the hearing aid, they produce an unpleasant noise or whistle detectable by the user and others in close proximity. Unfortunately, however, feedback oscillation is difficult to control because of the close proximity between the microphone and other components of the hearing aid, e.g. the amplifier.
Presently, two predominate methods exist to compensate for feedback signals in hearing aid systems. The first method involves using a filter to calculate the best set of filter coefficients for lowering the gain or power of the feedback signal at the offending frequency to prevent oscillation. This technique, however, suffers from the disadvantage of limiting the actual output power available for the hearing aid. In addition, it can also decrease the ability of the patient to clearly understand speech, especially when background noise is present, and/or the speech includes an accent.
The second method involves injecting a signal with the same behavior as the feedback signal only out of phase by about 180 degrees. The injected out of phase signal operates to cancel out the offending feedback signal. For instance, in one particular application of this method described in European Patent Application No. 0 415 677 a digital pseudo-noise signal is supplied to a digital filter and to a feedback path of the hearing aid. The noise signal provided to the digital filter and the feedback path is received at a summation point with the output of the summation point being provided as one input to a digital correlator, while the original noise signal is provided as the other input to the digital correlator. The individual delay stages of the digital correlator produce digital values that are used for adaptive optimization of the coefficients of the filter. This process causes continuous matching of the digital filter to the conditions of the feedback path for the hearing aid. In another application of this technique, a pulse generator is provided for feeding short individual pulses to the feedback signal path that are utilized to determine the impulse response of the feedback signal path. The impulse response is then used to measure the transfer function of the path and set the filter coefficients. The disadvantage of these techniques, however, is that they address the problem of feedback through reactive compensational responses, which helps, but does not solve the problem. A further disadvantage is a comparatively high cost of digital processing. For instance, one coefficient multiplication in the digital filter requires at least two more multiplications with variable factors for filter adaptation.